vy-p-tran/predicting-wine-quality

Predicting wine quality scores

Project overview

The goal of the project is to predict wine quality based on physicochemical data. The data comes from Portuguese Vinho Verde wine, including quality score, which was used as label for the model. Linear regression with lasso was performed to shrink the coefficients of the variables. The model results in RMSE of 0.76. The linear regression formula was used in this Shiny app to predict the wine's quality score.

• Modeling quality of wine.R: Codes for training and testing the model
• Shiny_app.R: Codes for Shiny app
• wine.csv: data set

Link to Shiny app

https://vyptran.shinyapps.io/Wine-Quality-Prediction/

Link to video

https://youtu.be/vxcyICBmnJw

Requirements

• For Shiny app: package Shiny
• For training and testing the model: packages dplyr, caret, and glmnet.
• All analyses were done with R (version 3.6.2)